I am struggling with a DOS condition that I believe is described in this
thread http://sourceforge.net/mailarchive/message.php?msg_id=30410146. An
interrupted negotiation between client and server can cause the nbd-server
parent process to fail to complete future connection attempts. Restarting
the nbd-server service will correct the problem but also severs the
connection for the other 30-45 clients that are connected leaving them with
unresponsive clients and squashfs errors. I believe this situation
is occurring when students inadvertently interrupt the boot process by
power cycling a client during boot or when a switch freezes or is
disconnected during a boot process. This occurs infrequently, on the order
of once a week across 4 servers and ~200 clients, but is disastrous for the
Ubuntu Precise incarnation of LTSP we upgraded to this year. My hope is
that I can return to spawning individual instances of nbd-server so that
this flaw in the forking model of nbd-server does not effect every user
attached to the system. I have tried variations on the old inetd setup
used in 10.04 but the named export convention and the change in port for
nbd-server seem to be thwarting me. Is anyone successfully running the
Precise version of nbd-server from inetd? If so, a look at your
configuration would be greatly appreciated.
--
Travis Borders
borde...@woodlandschools.org
Assistant Programmer/Web Developer
Woodland School District
360-841-2734
------------------------------------------------------------------------------
Everyone hates slow websites. So do we.
Make your web apps faster with AppDynamics
Download AppDynamics Lite for free today:
http://p.sf.net/sfu/appdyn_d2d_feb
_____________________________________________________________________
Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto:
https://lists.sourceforge.net/lists/listinfo/ltsp-discuss
For additional LTSP help, try #ltsp channel on irc.freenode.net
